OpenAI has shut down a recently introduced ‘Make this chat discoverable’ feature in ChatGPT after concerns were raised about the unintended exposure of personal conversations. The feature, launched as a limited experiment, allowed users to share specific chats and make them searchable on platforms like Google.
ChatGPT's public sharing feature has been removed
Although the tool required multiple opt-in steps and anonymised the content, several incidents revealed that sensitive user data was still making its way into public search results.
Why was the feature removed?
Dane Stuckey, OpenAI’s Chief Information Security Officer, confirmed the removal on X (formerly Twitter), stating: “Ultimately, we think this feature introduced too many opportunities for folks to accidentally share things they didn’t intend to, so we’re removing the option.”
The decision followed a report from Fast Company, which noted that search engines had started indexing user-shared conversations—some of which included discussions around mental health, workplace issues, and personal experiences.
Shared chats indexed by Google
Despite the feature requiring manual selection and confirmation before publishing, many users misunderstood its implications. Although usernames were hidden, the content itself often contained enough detail to pose privacy risks.
This sparked concern across social platforms, where users shared examples of private-sounding conversations appearing in Google Search.
OpenAI starts delisting and cleanup
OpenAI has confirmed that:
- The feature is being disabled on all user accounts.
- The company is working with Google and other search engines to delist indexed pages.
- Delisting is expected to be completed by Friday morning.
OpenAI further emphasised that security and privacy remain a top priority, and it will continue refining its tools with this focus in mind.
Transparency vs. privacy in AI tools
The now-retired feature was initially introduced to help users share useful or educational chatbot exchanges with a wider audience. However, the incident underlines the complexity of ensuring safety and privacy in AI-powered tools, especially when users increasingly rely on them for both personal and professional conversations.
